Seediscussions,stats,andauthorprofilesforthispublicationat:http://www.researchgate.net/publication/24299263

ExperimentsWithFixedandAdaptiveHerschel-QuinckeWaveguidesonthePrattandWhitneyJT15DEngineARTICLEAPRIL2002Source:NTRS

DOWNLOADS28

VIEWS87

3AUTHORS,INCLUDING:

RicardoA.BurdissoVirginiaPolytechnicInstituteandStateUniv123PUBLICATIONS730CITATIONS

SEEPROFILE

CarlH.GerholdNationalInstituteofAerospace41PUBLICATIONS130CITATIONS

SEEPROFILE

Availablefrom:CarlH.GerholdRetrievedon:08July2015

NASA/CR-2002-211430

Experiments With Fixed and AdaptiveHerschel-Quincke Waveguides on the Prattand Whitney JT15D Engine

Jerome P. Smith and Ricardo A. BurdissoVirginia Polytechnic Institute and State UniversityBlacksburg, Virginia

March 2002

The NASA STI Program Office... in Profile

Since its founding, NASA has been dedicatedto the advancement of aeronautics and spacescience. The NASA Scientific and Technical

Information (STI) Program Office plays a keypart in helping NASA maintain thisimportant role.

The NASA STI Program Office is operated byLangley Research Center, the lead center forNASA's scientific and technical information.

The NASA STI Program Office providesaccess to the NASA STI Database, thelargest collection of aeronautical and spacescience STI in the world. The Program Officeis also NASA's institutional mechanism for

disseminating the results of its research anddevelopment activities. These results arepublished by NASA in the NASA STI ReportSeries, which includes the following reporttypes:

TECHNICAL PUBLICATION. Reports ofcompleted research or a major significantphase of research that present the resultsof NASA programs and include extensivedata or theoretical analysis. Includescompilations of significant scientific andtechnical data and information deemed

to be of continuing reference value. NASAcounterpart of peer-reviewed formalprofessional papers, but having lessstringent limitations on manuscriptlength and extent of graphicpresentations.

TECHNICAL MEMORANDUM.

Scientific and technical findings that arepreliminary or of specialized interest,e.g., quick release reports, workingpapers, and bibliographies that containminimal annotation. Does not contain

extensive analysis.

CONTRACTOR REPORT. Scientific and

technical findings by NASA-sponsoredcontractors and grantees.

CONFERENCE PUBLICATION.

Collected papers from scientific andtechnical conferences, symposia,seminars, or other meetings sponsored orco-sponsored by NASA.

SPECIAL PUBLICATION. Scientific,technical, or historical information fromNASA programs, projects, and missions,often concerned with subjects havingsubstantial public interest.

TECHNICAL TRANSLATION. English-language translations of foreign scientificand technical material pertinent toNASA's mission.

Specialized services that complement theSTI Program Office's diverse offerings includecreating custom thesauri, building customizeddatabases, organizing and publishingresearch results.., even providing videos.

For more information about the NASA STI

Program Office, see the following:

Access the NASA STI Program HomePage at http:]]www.sti.nasa.gov

Email your question via the Internet tohelp@sti.nasa.gov

Fax your question to the NASA STIHelp Desk at (301) 621-0134

Telephone the NASA STI Help Desk at(301) 621-0390

Write to:

NASA STI Help DeskNASA Center for AeroSpace Information7121 Standard DriveHanover, MD 21076-1320

NASA / CR-2002-211430

Experiments With Fixed and AdaptiveHerschel-Quincke Waveguides on the Prattand Whitney JT15D Engine

Jerome P. Smith and Ricardo A. Burdisso

Virginia Polytechnic Institute and State UniversityBlacksburg, Virginia

National Aeronautics andSpace Administration

Langley Research CenterHampton, Virginia 23681-2199

Prepared for Langley Research Centerunder Grant NAG1-2137

March 2002

Available from:

NASA Center for AeroSpace Information (CASI)7121 Standard DriveHanover, MD 21076-1320(301) 621-0390

National Technical Information Service (NTIS)5285 Port Royal RoadSpringfield, VA 22161-2171(703) 605-6000

TABLE OF CONTENT

TABLE OF CONTENT ................................................................................................ 1ABSTRACT .................................................................................................................. 31. INTRODUCTION ................................................................................................. 52. EXPERIMENTS WITH FIXED HQ TUBES ...................................................... 7

2.1 THE JT 15D ENGINE AND TEST CELL ....................................................... 7

2.2 THE HQ INLET ............................................................................................. 72.3 MEASUREMENT FACILITIES ..................................................................... 9

2.4 EXPERIMENTAL RESULTS ...................................................................... 10

3. EXPERIMENTS WITH ADAPTIVE HQ TUBES ON THE JT15D ENGINE 143.1 ADAPTIVE-FLAP HQ-TUBE SYSTEM ..................................................... 14

3.2 EXPERIMENTS USING THE ADAPTIVE-FLAP HQ-TUBE ..................... 18

3.3 ADAPTIVE-LENGTH HQ-TUBE SYSTEM ............................................... 22

3.4 EXPERIMENTS USING THE ADAPTIVE-LENGTH HQ-TUBE ............... 253.5 ADAPTIVE CONTROL SYSTEM ............................................................... 26

4. CONCLUSIONS ................................................................................................. 345. RECOMMENDATIONS FOR FUTURE RESEARCH .................................... 34ACKNOWLEDGEMENTS ........................................................................................ 35REFERENCES ........................................................................................................... 35

ABSTRACT

This report presents the key results obtained by the Vibration and AcousticsLaboratories at Virginia Tech over the period from January 1999 to December 2000 onthe project "Investigation of an Adaptive Herschel-Quincke Tube Concept for theReduction of Tonal and Broadband Noise From Turbofan Engines" funded by NASALangley Research Center. The Herschel-Quincke (HQ) tube concept is a developingtechnique that consists of installing circumferential arrays of HQ tubes around the inlet ofa turbofan engine. This research is a continuation of previous efforts in which the HQconcept was preliminarily validated on the JT15D engine [1].

This final project report is organized in three separate reports. The research presentedin these reports summarizes both analytical and experimental investigations of the HQconcept for reducing turbofan radiated inlet noise. The analytical part of the projectinvolves two different three-dimensional modeling techniques to provide prediction anddesign guidelines for the application of the HQ-concept to turbofan engine inlets. First,an infinite-duct model was developed and used to provide insight into the attenuationmechanisms of the HQ systems and design strategies. Results from this analytical modelshow the effectiveness of the technique and are presented here. Second, the NASA-developed TBIEM3D code was modified to allow numerical modeling of HQ systems.This model allows for the investigation of the HQ system when combined within apassive liner. The experimental part of this work includes data for "fixed" HQ tubes onthe JT15D engine with different inlet acoustic modal content than previously tested.Experimental results for fixed HQ tubes on a full-scale Honeywell TFE731-60 engine arealso presented. Also included here is the first set of results of an experimentalinvestigation into adaptive HQ configuration on the JT15D engine. The parameters of theHQ tubes are changed to optimize the attenuation as the engine speed is changed.

The first report presents the analytical modeling and simulation results. The secondreport describes the experimental results with both fixed and adaptive HQ-tubes on theJT15D engine. Finally, the third report describes the most important results with fixedtubes on the Honeywell TFE731-60 engine. The three parts of this final report arewritten such that each part is a complete and separate document that can be reviewedindependently from the others.

1. INTRODUCTION

The Herschel-Quincke(HQ) tube conceptconsistsof installing circumferentialarraysof HQ tubesaroundtheinlet and/orthe by-passductof a turbofanengine.Theapplicationof HQ tubes to turbofan engineinlet noise is a developingtechniqueoriginally pioneeredat Virginia Tech. The researchpresentedin this report is acontinuationof previouseffortsin whichtheHQ conceptwaspreliminarilyvalidatedonthe JT15D engine [1]. The accomplishmentsof the previousresearchefforts aresummarizedin anearlierreport[ 1]. Themainpreviousachievementsinclude:

Experimentalresultson theJT15Dengineinlet demonstratedBPFtonepowerattenuationof up to 8 dBwith fixedarraysof HQ tubes.

TheHQ tubeconceptalsoprovidessignificantattenuationof the broadbandcomponent(- 3 dBpowerreductionover0-3200Hz band.)

An initial analyticalmodelwasdevelopedto investigatethe noisecontrolmechanismsof the HQ tube concept and to guide in the design ofexperiments.

An overviewof thetasksinvolvedin thisprojectis shownin Figure1.1.Theprojecthas analytical and experimentalcomponents.The analytical part involves thedevelopmentof two modelingtools for theHQ-tubeconceptappliedto turbofanengineinlets.The experimentalpart consistsof validatingthe approachin two engines,i.e.Pratt&Whitney JT15D and Honeywell TFE731-60engines, for various HQ-tubeconfigurations.Themainobjectivesof thiscontinuingresearcheffort are:

To further develop modelingtechniquesfor the design,prediction,andoptimizationof the Herschel-Quincke(HQ) tubeconceptfor applicationtoturbofanenginenoise.

To experimentallyinvestigateboth fixedandadaptiveHQ-systemsfor usefulreductionof turbofaninlet noisewith realisticcomponentson a runningturbofanengine.

Thefinal reportis organizedin three parts devoted to the various components of theresearch endeavor. This report corresponds to Part II, which describes the experimentalwork performed on the Pratt & Whitney JT15D engine at Virginia Tech. Section 2 of thereport presents tests using fixed HQ-tubes on the JT15D without the upstream rods usedin previous experiments [1]. Removing the rods results in a different set of dominantmodes in the inlet. Section 3 contains the first set of results using adaptive HQ tubes onthe JT15D engine. Two adaptation mechanisms are described and tested on the engine. Acontrol system was also developed to adjust the parameter of the HQ-tubes to maximizethe attenuation of the BPF tone as the engine speed was changed.

ANALYTICAL WORK EXPERIMENTALWORK

VPI InfiniteDuctModel

Extendmodel. Studytoneandbroadbandcontrol. Investigatenoisecontrolmechanisms. Optimization. Designexperiments.

_Engine Experiments_

FixedTubes y

JT15D engine 1 and 2 arrays._" m=l and 1.t=0,1,2,3_, m=5 and l.t=O,1

Honeywell TFE731-60 Engine._" m=2,-8,12 and kt=O, .... 5

_'_

Model y

Implement HQ-tube modeling toTBIEM3D

Combined Liner-HQ system. Study forward and aft. radiation. System optimization. Modeling of high-bypass engines.

_Engine Experiments_

Adaptive Tubes _

Investigate tube adaptation mechanisms. Evaluate adaptive tubes on JT15D engine. Implement adaptive control system. Demonstrate adaptive HQ system on JT15D

engine.

Figure 1.1: Overview of project tasks.

2. EXPERIMENTS WITH FIXED HQ TUBES

This section presents the experimental setup and results for the HQ concept with fixedtubes on the inlet noise radiated by the JT15D turbofan engine at Virginia Tech. Thefixed-tube experimental research goals for this year were to demonstrate the HQ concepton the JT15D engine without the exciter rods installed upstream of the fan. Removingthese rods resulted in a different acoustic inlet modal configuration than previously testedwith the rods.

2.1 THE JT15D ENGINE AND TEST CELL

The engine used for this research project is a Pratt and Whitney JT15D turbofanengine. This engine has been used in previous NASA studies and has been usedextensively for research in active noise control methods applied to turbofan engines [2-5].It is a twin spool turbofan engine with a full length bypass duct and a maximum bypassratio of 2.7. There is a single-stage axial flow fan with 28 blades and a centrifugal highpressure compressor with 16 full vanes and 16 splitter vanes. There are no inlet vanesand the diameter at the fan stage location is 0.53 m (20.8 in). Experimental results wereobtained by operating the engine at various speeds near the idle condition whichcorresponds to a fan speed range of approximately 5250 rpm, yielding a blade passagefrequency (BPF) ranging from 2225-2520 Hz. Near this condition, the inlet intake flowspeed is about 42.5 m/s which yields a Mach number of M=0.12. The engine is equippedwith an inlet inflow control device (ICD) constructed at Virginia Tech from a NASAdesign [6]. The purpose of the ICD is to minimize the spurious effects of ground testingon acoustic measurements by breaking up incoming vortices. The maximum diameter ofthe ICD is 2.1 times the engine inlet diameter.

In previous tests, in order to enhance the tonal nature of the inlet radiated sound andto excite the m =1 mode to dominance, a set of 27 exciter rods were mounted upstream ofthe fan stage [1]. In the tests presented here, the rods were removed to provide a differentmodal configuration.

The engine test cell consists of two chambers, with the forward section consisting of asemi-anechoic chamber to simulate free field conditions. One wall of the semi-anechoicchamber is open to the atmosphere for engine intake air.

2.2 THE HQ INLET

In order to facilitate the rapid installation and removal of the HQ tubes on the engineinlet, a compact inlet section was constructed, which could be configured as either a rigidwall or with HQ elements. The inlet consists of a perforated mesh cylindrical skeleton,supported at each end by two circular plate rings and in the middle by four rectangularbeams located geometrically 90 apart. The HQ elements could then be mounted behindthe perforated mesh cylinder. A hard-wall inlet could be implemented by mountingsections of a hard, rigid material behind the mesh skeleton. The hard material used here

was6.3 mm (0.25 in) thick ABS plastic. The inner diameter of the inlet was 0.53 m (20.8in) in order to match the diameter of the engine at the fan stage where the inlet was to bemounted. The length of the inlet in the axial direction was 0.46 m (18 in), including thetwo 6 mm (0.25 in) thick plate rings at each end of the inlet to allow attachment of theinlet to the engine at one end and the attachment of the ICD to the inlet at the other. Inthese experiments, one or two arrays of HQ tubes are mounted circumferentially aroundthe cylindrical perforated mesh inlet of the turbofan jet engine. The surface area of theinlet section where the tubes were not attached was configured as a rigid wall. The inlethas a length/diameter ratio L/D _ 1, which is typical of turbofan engines. All of the tubesare axially-oriented (i.e., parallel to the engine axis) as shown in the configurationschematic in Figure 2.1. Figures 2.2(a) and 2.2(b) show pictures of the engine inletconfigured as a rigid wall and with two arrays of HQ tubes, respectively. In Figure 2.2(b)the bottom panel is left off so that the mesh screen cylinder is visible.

Array #1 Array #2 NOTE:20 Tubes 16 Tubes 0 rods1. 1, i" I.D.

JT15D

Fan

17.5Scale ~ 1:7

Figure 2.1: Schematic of experimental configuration for fixed HQ tubes on theJT15D without rods.

The tubes and panels were both constructed of black ABS plastic. The tubedimensions were designed so that their second resonance occurs in the vicinity of the firstengine BPF tone near 2320 Hz, and each have a diameter of 3.8 cm (1.5 in) and aneffective bypass length of 11.2 cm (4.4 in). The case shown in Figure 2.1 is for the firstarray (away from the fan) containing a total of 20 tubes, and the second array (closest tothe engine) containing 32 tubes. However, the second array can also be configured with16 tubes by simply blocking half of them. When both arrays were implemented

simultaneously,thetubesoccupiedlessthan8.1%of thetotal surfaceareaof the inlet.The adaptive HQ tube arrays were also mounted in a similar fashion. The mechanismsand dimensions of the adaptive HQ tubes will be presented in the experimental resultssection.

(a)

(b)

Figure 2.2: Pictures of experimental configuration for (a) hard wall, (b) fixed HQ-tubes(2 arrays of 20 and 32 tubes) on the JT15D without rods.

2.3 MEASUREMENT FACILITIES

The acoustic field of the JT15D engine was monitored with an array of 31 farfieldmicrophones positioned in the horizontal plane passing through the centerline of theengine. The microphones were spaced along an arc of radius 1.6 m (63 in) at 6 increments to obtain the acoustic directivity from -90 to 90 in the horizontal plane,(where 0 is along the engine axis). These microphones were used to evaluate the effectsof the HQ tubes on the noise radiated by the engine.

2.4 EXPERIMENTAL RESULTS

Presented here are the key experimental results obtained with one and two arrays ofHerschel-Quincke (HQ) tubes installed on the JT15D engine inlet, with the engineconfigured without excitation rods. All previous tests with the HQ tubes were performedwith 27 rods located upstream of the fan [1]. These tests performed without the rodsshow the effect of the HQ tubes for the inlet with a different modal configuration. Bothbroadband and tonal data were obtained for a hard-walled inlet and for the inletconfigured with one and two arrays of HQ tubes. Figure 2.1 shows a schematic of theengine configuration with both arrays of HQ tubes. When testing a single array, onlyArray 1 was used which contained 20 tubes. When implementing two arrays, Array 2was added which contained 16 HQ tubes. As in previous tests, all tubes were identicaland designed for their second resonance to occur near the BPF of 2320 Hz. The enginespeed was varied to obtain data at several different BPFs ranging from 2260 to 2430 Hz.

The cut-on frequencies for the excited modes in the JT15D inlet with and without the27 rods installed, respectively, are shown in Table 2.1. For the case with 27 rods, the 28fan blades interact with the 27 rods to excite to dominance the m=l circumferentialmode. As shown in the table for the case with 27 rods, there are three radial modes cut-on below 2381 Hz, and the fourth radial mode cuts on above that frequency. For the casewithout rods, the only interactions that exist are between the fan blades and the statorvanes present in the engine. There are 66 stator vanes in the bypass region and 33 statorvanes on the core. The 28 fan blades interact with the 66 bypass stator vanes to excite them=28 circumferential modes at BPF. However, all modes with circumferential orderm=28 are cut-off below a BPF of about 6200 Hz, and thus these are evanescent modes,i.e. fan-bypass stator interaction is cut-off at the running speeds tested here. The 28 fanblades also interact with the 33 core vanes to excite m =5 modes at BPF. As shown in thetable for the case with 0 rods, two radial modes with circumferential order m=5 are cuton above 2162 Hz. Thus, for the range of running speeds tested here without the rods,there appear to be two modes cut-on, the (5,0) and (5,1). Note that the interaction of thefan with the stators also takes place when the rods are installed, but the excitation of them=l modes due to the rods is significantly stronger than the excitation of the m=5 modesdue to the core stators.

Table 2.1: Cut-off frequencies of JT 15D Inlet Modes.27 Exciter Rod 0 Exciter Rods

Mode(1,0)

Cut-off Freq. [Hz]374

Mode(5,0)

Cut-offFreq. [Hz]1306

(1,1) 1084 (5,1) 21621737(1,2)

(1,3) 2381(5,2)

(1,4) 3021

2846

The acoustic power at the BPF tone was determined for the total sector (0 to 90 )and the sector toward the sidelines from 50 to 90 for the hard-walled inlet and the inlet

10

_th oneandtwo arraysof HQ ttthes_FigureZ3(_) shows_e BPF tortspowerredt_etionov_ the 0 to 90 sector for the ease without rods for v_ous BPF tone frequenciestested. For the sake of comparison, Figure 2.3(b) _hows the BPF power reduction for thesame inlet e.o_gurations _ith the rods installed. It shoed be noted that the power of theBPF tone for the hard.walled i_et was about: 15 dB higher with the rods installed, as itwill be shown later, For the engine without rods, _e HQ tubes show good reduet:on ofthe BPF tone power over the frequency range tested, The addition of the second arrayresalts in an average increase of _proximately 2 dB in the achieved reduction of the BPF

tone power. The most reduction was achieved with _o arrays _t a BPF of 2264 Hz (thelowest _nency on the range t_ted)with a total BPF power reduction of 4,9 _, Thepower reduction over the sector from 50 _ }o 90 _ was 5o4 dB. From the data, it appears asthough the optimal BPF for both one grid two HQ tube arrays may be below the lowest_uer_ey in the range tested here,

(a)

(b)

v

8-

_2

(a) 0 Rod_ (m-_5)

..................................................................................... .,...,.,..,.w_... ,... ,...,,..,. _ ...........................................................

2300 _50 24f_0 245_ ZS@) ZSNBPF Tone Frequency [Hzl

4

2250 2300 2350 2400 2450 2500 255(BPF Tone Freqneney [Hzl

F_gt_re 2.3: Total power reduction a_ BPF torie over 0 to 90 seetor with fixed HQ tube_for (a) 0 exe_ter rods arid (b) 27 exciter rods.

,,r,rrrrrr,rrrrrrrrrrrr r

The broadband power spectra for the 0 to 90 sector without and with rods installedare shown in Figures 2.4(a) and 4(b), respectively. These figures show that the presenceof the rods increased the BPF tone power - 15 dB while the broadband components didnot change appreciably. The broadband content shows reduction over most of thefrequency range with the most significant levels of reduction in the vicinity of the first(-1200 Hz) and second (-2400 Hz) resonances of the tubes for both with and withoutrods cases. However, the broadband attenuation is better for the case with rods. In figure2.4a, broadband power reductions exceed 6 dB with 2 arrays of HQ tubes at frequenciesin the vicinity of the first resonance. The overall broadband power reduction over thetotal sector (excluding the tone) was 1.7 dB with 2 arrays and 0.9 dB with one array.These overall reductions would only increase by about 0.1 dB if the reduction at the tonewas included in the calculation of the overall reduction. Thus, the tone reduction doesnot contribute significantly to the overall reduction when the rods are removed. Theoverall broadband power reduction over the sector from 50 to 90 (excluding the tone)was 2.2 dB with 2 arrays and 1.3 dB with one array, and again these overall reductionswould only increase by 0.1 dB if the reduction at the tone was included in the overallreduction.

It is evident from the experimental data that the HQ tube concept is effective forreducing the inlet noise radiated from the JT 15D engine without the presence of exciterrods. Without the rods, two radial modes with circumferential order m=5 are excited.The HQ concept results in both tonal and broadband noise reduction for this modalconfiguration. Two HQ arrays yielded a 4.9 dB reduction in the BPF tone power and anoverall broadband reduction of 1.7 dB. As in previous experiments with the rods, twoarrays of tubes are shown to increase the reduction achieved at both the tone and in thebroadband levels.

12

(a)

(b)

120

t05_

_OO 9S

8_

80:

- Hard Wa_U._-_-_-_-_-1 Away (20 Tub_s)...... 2 Arr_ (20 & _6 _bcs)

iO00 _500 2000 2500 3000

Frequency (Hz)

120 ]

._. 1t5 ]

- t10595-

90.-

80*gO0 1000 _500 2000 2500 3000

_equen_ (Hz)

Figure 2A: Broadband pow_x reductionwith fi.xe_ HQ t_be_ (a) no _ (b) 27 rod_.

13

3. EXPERIMENTS WITH ADAPTIVE HQ TUBES ON THE JT15D ENGINE

In this section, the setup and results of the preliminary investigation into the potentialof using adaptive HQ tubes is presented. It is clear that although the tubes are veryeffective over a range of BPFs, there is an optimal BPF at which the best attenuationoccurs. This implies that in order to optimize the attenuation effect of the tubes over awide range of BPF tones, the tubes should be adapted to "track" the BPF as it changes.The optimal attenuation frequencies depend on the geometry of the HQ tubes. Thus, thetube properties can be adjusted in real time for optimal tube performance as the BPF tonechanges with engine conditions. The properties of the tube such as the length, cross-sectional area, and so forth can be modified in real time to provide optimal reduction ofthe inlet fan tone noise over the desired sectors in the far-field. The adaptation of the tubedynamics can be implemented with a control scheme (i.e. feedback, feedforward, etc.) asdepicted in the schematic in Figure 3.1. The error signal used by the controller will beobtained from either far-field or inlet mounted microphones. The signals from thesemicrophones will be processed to generate a signal proportional to the acoustic powerradiated by the fan. The controller will then adjust the tube properties to minimize theobserved radiated power. Since the optimal frequency of reduction for a set of HQ tubesis related to the HQ tube resonant frequencies, an effective actuation mechanism isexpected to be one that effectively and linearly changes the resonant frequencies of theHQ tube. It is important to remark that the purpose here is not to develop a practicaladaptation mechanism but rather investigate the potential of adaptive HQ-tube systems.

Sensors toestimate power

_ Control

I _ -_Systeml_k. l I I _ HQ-system

t,"J._ f__ , .

..- . Inlet _ Engine axis

Figure 3.1: Schematic of adaptive HQ tube system applied to turbofan engine inlet.

3.1 ADAPTIVE-FLAP HQ-TUBE SYSTEM

Figure 3.2 shows one of the adaptation mechanisms proposed and preliminarilyinvestigated. An internal "throttle-plate" flap is introduced into the tube and changes theresonant frequencies of the HQ tube when the angle

end,which was left open. The speaker was driven with white-noise and the transferfunction between the speaker input and the microphone was obtained for various flapangles. Note that this test configuration does not accurately determine the absolutefrequency values of the open-open tube resonances due to the placement of the speaker,but it provides a simple test in which the actuation mechanism can be evaluated in termsof the change in the resonant frequencies.

RotationalFlap/

Flap a = 0 Flap c_= 90 I

Figure 3.2: Schematic showing adaptive-flap HQ tube mechanism

Figure 3.3 is a plot of the frequency response between the speaker and themicrophone for various values of the flap angle ix. The first and second resonancefrequencies of the HQ tube with the flap are indicated by the arrows. Note that the othertwo peaks in the FRF near 600 Hz and 1300 Hz are speaker effects as they weredetermined by testing the speaker alone. It is clear from this plot that the secondresonance frequency of the tube increases with the flap angle ix. However, it is also clearthat the shape of the response around the resonant frequency changes with the flap angleas well. It is believed that the most desirable resonance for optimal reduction with theHQ tube is a sharp, well-defined single peak, the best example being the response curvefor ct = 30 . Thus, it appears that desirable resonant behavior will occur when the flapangle is near tx = 30 and that the response may have some non-linear effects as ct getstoo large or two small. For the flap at ct = 0 , the HQ-tube is theoretically transformedinto two quarter-wave tubes of half the length of the HQ-tube. Thus, the first resonancefrequency of the quarter-wave tube should be half the second resonance frequency of theHQ-tube. However, it is important to note that the flap did not perfectly sealed the tube toform two quarter-wave tubes, and thus there were leaks between the flap and the insidewall of the tube.

It should also be noted that changing the flap angle does not significantly change thefirst HQ resonant frequency near 1000 Hz. This is due to the fact that the flap is locatedat the center of the HQ tube, where for the odd resonant frequencies there is a pressureanti-node (i.e. pressure is maximum and the particle is zero). Because the flap leads tochanges in the tube cross-sectional area, it cannot affect the kinetic energy of the oddmodes. On the other hand because the even modes have a pressure node at the tube'scenter (i.e. maximum particle velocity), the kinetic energy increases as the flap reducesthe cross sectional area and the resonance frequency drops. The potential of affectingonly even modes could be a desirable approach. For example, if the 3rd tube resonance is

15

targeted to control the BPF tone at a high engine power setting; adjus_ng the 2 _'_resonance of the tube to be optimum at controlling the BPF tone at: a tower engine _wersetting may be req_tired. ,_other potemial application is if the optimal frequencies of_uction for the first HQ resonar_ce _sh to be kept fixed (cog., for low_freque_cybroadband noise), while onty the second resonance corresponding to the chm_gigg BPF isto be changed.

+

_0 s*

=30 _.....................a= _5

-- a=75_

500 1(K_ .1500 20_X_ 2500

Frequency (Hz)

Figure 3,3, FRF she _mg adapt_ve_flap HQ tube resonaat _uency change with flapaagle.

Tweaty tubes were. constructed with the adaptive flap mech_s_r_ instMied on theJTI 5D, and the a.tten:t_atioa effeets evaluated at: :different BPF frequencies far flap _gle_ranging _om 0 (completely shut} to 90 _ (open or hor_zontaL) A stepper-motor systemwa_ implemented to control the flap aagles with a 2 resolution. Figm'e 3_4 shows aschematic of the inlet as confib_ with the adaptive HQ tube array. Also i_staIled on

effect_ of the adaptive-- array were tested alone a_d _a additio:a to the sir_gle fixed _'array,Figure 3.5 contains a pict_re showing the adaptive array inlet mounted art the _i 5D arida close-up of the adaptation mechanism with a stepper motor_ A total of fern" stepper

" . _ , ._ _ t, .......motors were used, with one motor m_.tmted o_ each ot _e _bur quar_er-p_els to cor_trolthe flaps oa their respective panels. AlI tb_ motors were controlled by a single s_pper-metar eoatrotler which s_pptied the same actuation sig_N to each mot_r_ The motor_were connected to the tube flap axes by a direct-drive ctvaJrt as showt_ in the pietttre. Therotating .fl_ axN of each tabe oa the sa_ae pa_et were c_aaect_ wi_h uNve_al joiats sothat NI of the flaps on the same partel moved together.

Adaptive Array Fixed Array20 Tubes 16 Tubes 27 rods

Fan

JT15D

14.5

Inlet22" I.D.

17.5Scale ~ 1:7

Figure 3.4:. Schematic of experimental configuration for adaptive-flap HQ tubes on theJT15D.

Figure 3.5: Pictures showing experimental configuration for adaptive-flap HQ tubes onthe JT 15D.

17

3.2 EXPERIMENTS USING THE ADAPTIVE-FLAP HQ-TUBE

The adaptive-flap HQ-tube system shown in figure 3.5 was tested on the JT15Dengine. Figure 3.6 shows the total BPF power reduction over the range of flap angles forthree different BPF frequencies with the single adaptive array. Starting with the BPF of2360 Hz (the blue curve) a total of 6 dB reduction is achieved with an optimal flap angleof x=30 . As seen by moving downward to the magenta curve, if the engine speed wereincreased to a BPF of 2560 Hz, the BPF reduction would drop to about 3.2 dB. But, aspredicted by the flap characterization, opening the flap angle to cx=50 results in anincrease of the BPF reduction to about 5 dB at the higher BPF. Now, starting again at theblue curve with cx=30 , if the engine speed were to decrease to result in a BPF of 2280Hz, the reduction would drop to about 4 dB. The earlier characterization of the flapmechanism would imply that the flap angle needs to be additionally closed to result inoptimal reduction for a lower BPF, but as seen in the figure, closing the flap does notresult in an increase in the BPF reduction. The best reduction achieved at a BPF of 2280Hz was about 4.8 dB, with the flap angle completely closed. Thus, it appears that the flapmechanism does not perform well for the extremes flap positions, i.e. c_-0 and cx-90 .

The effects of the adaptive-flap HQ array on the BPF power reduction with theaddition of the second fixed array for four different BPFs are shown in Figure 3.7. In asimilar fashion as in the previous figure, starting at the BPF of 2340 Hz (the dark blueline) at the optimal flap angle of _=30 , the BPF reduction is seen to be almost 8 dB. Ifthe BPF were to change up to 2520 Hz, the reduction would drop to only 4.5 dB. If theflap angles were now opened to cz=50 , the reduction would increase to about 7.3 dB.Unfortunately, no flap angles existed for regaining such a significant amount of BPFpower reduction when changing the engine speed above and below this BPF range.

It should be noted that at several BPF test frequencies, the effect of the flap angle onthe total BPF power reduction did not follow an observable trend with the change in flapangle _x. The initial characterization testing of the adaptive flaps did show the potentialof this particular adaptation concept for increasing the reduction obtained with the HQtubes over a range of running speeds. However, the range of frequencies obtainable withthe flap mechanism was somewhat limited.

18

87

A

= 5O4

3

{20

gure 3.6.. Power reduction at the BPF to_e versus flap angle for several BPFs usingthe single adaptive-flap array.

Figure 3.7.. Power reduction at _he BPF tone versus fl,tp angle for several BPFs uzmg e single _apfive-.fi_ array in combination with the fixed array.

_e effect of the ad_tive flap on. the inlet radiated broadband rtoise was alsoinvestigated_ Figure 3.8 sbows the to_ A-weighted power spectra from 0 to 3200 Hz fbrthe hard-wNl inlet arid for the, _ingI.e adaptive-flap HQ array with flap angles of a_50 _and a=80 . Figure 3_9 contains the same plot for the case with the single adaptive-flaparray and the fixed, array with t 6 tubes both on the intet_ It is evident from these plotsthat the flap angle clearly has an effect oa the ftequertcy of optimal red_aeti.on due m thesecond resor_anee of the tubes (near 2400 Hz)_ while the optimal frequer, cy of reduction

19

for the first resonm]ee (near 1300 HZ) remains asseatially unchartged_ It appears that theoptimal .frequency of reduction due to the sec, ot_d resonar_ce is ne_xr 2000 Hz with the flapangle at t_SO _, _d uP sear 2600 Hz with o_=80 + for both the single adaptive+lisp arrayalone and the single adaptive-flap array together with the fixed HQ array, Also ele_ is_e significant h_crease in the overalI reduction with _he two _rays of robes as comparedto the single adaptive array anne.

85

80

75NNI

.............. Hard Watl ]

1--a_50o

cleo i_fJ(l 20{)0 2500 30_

Figure 3.8:Total A-weighted power spectra using the single adaptive-flap array.

tl,9 +

105 +

i+8s

i ..........................i!1_...... _--+_o I

800 1000 t500 290e 2500 3000

Fignre 3+9:Total A-weighted pe_e_ spectra using the _ingte adaptive-flap am_ay mcombination with the fixed arra_

2O

general the optmlat frequency of reductmr_ m the broadband noise content wa_seen to incre_e with increasing flap ang!e, as predicted with the charactefizafio_resonance tests. This effect is more obvious_'_' in the plots of Fig*are 3. i 0 (g) a_d _) whichshow the total broadband power red_ction versu_ .frequency and fl_ angle t_ for thesinNe array of 20 adaptive robes and the: two-array ease. The first optimal frequency ofreduction appears tobe tmaffected by _he change in flap a_Ne. However, the frequencyvmation of the second optimal _uency of r_uction can _ seen as a near %e- *_;hape,ie,_ does not change significantly _th flap angle at the low and high a_Nles (_m _=4) _to _=20 and from _._;0 _ m :a-_90), but clhanges atmost li_early win flap anglesbe_'een a=25 _ and c,=55L Thus, it: seems tha_ the :flap angle results in a eon._roliableeffect only over a relatively smatl flap a_gle r_ge, m a relatively small_uency range R is noted here that _e ab:setm_ level of broadband noise reduction Ndnot change si_ificanfly with the flap angle, as shown in Figures 3.11(a) and (b) showk_gthe A-weighted broadband power reduetioa from O to 3200 Hz excluding the tone versuafl_g_angle for Ne one ea_d two array cases, respectively.

Figure 3A0:Totat power re&acfiot_ vers_s y and. flap angie (a) single _aptive-flap areay (b) single adaptive_Ilap array m eombhaation with _e fixed arra}_

21

(3)

o 5 _o _ 2o z_ _ _5 4o 45 5o 55 6o 65 70 75 80 8s .,_

3.0--(b)

_2,.o iI,.5,

_ 1,0..

i t0 5 10 I5 N _ 30 35 40 45 50 55 60 65 70 75 80 85 90

F_gure 3A1 :To_at A_weighted broadband power reduction 0_3200 Hz (:no torte) versusflap angle (a) single adapt_ e-flap arra3_5 (b) single adaptive-flap _ay m eombmatto_

with the fixed array.

3,3 ADAPTIVE-LENGTH HQ-TUBE SYSTEM

Due m the drawbacks stated concerning {he adaptive-flap HQ array adaptation_echniqtte, ano*_er adaptatkm approach involving ehanging the length of the tubes wasinvestigate, It was expected _ha.t changing the tube length wo_Id provide.a more lhxea,"adaptat_ivm mechanism. The first design of the "_aptive-iength mec.has_igm is sho,_m mFiga_re 3A2, Both tube ends are attached to the inlet thmttgh an expendabie rubbersection, The tube Ieng_h is _hen controlled by the rota:tion of a screw into a _.ut fixed tothe HQ-tube, A stepper motor then drives the serew, By turning the _rew, the robe couldbe pulled up and down, a_Iowir_g the effective length of the: tube m change. Seve.mImechanisms for ehanNng the tube length we_ investigated; _e method ehesen was

x_4.

based on ease of mechanization. The advantage of the adaptive length mechanism is thatmoving the tube a height "d" results in a change in the effective length equal to 2d. Thesecond system consists of tubes of changeable length. The tube lengths rangeapproximately from 11.5 to 13.3 cm.

stepper motor

expandablesection

Figure 3.12: Schematic showing prototype for adaptive-length HQ tube mechanism.

The adaptive-length tube was characterized in the same manner as the adaptive-flap tubeexplained earlier, i.e., by placing a speaker at one end and a microphone at the other, andmeasuring the transfer function response with a white noise input to the speaker. Thefrequency responses showing the resonant frequency change for a variable change inlength A are shown in Figure 3.13. Both the first and second resonances show clear,sharp resonant behavior and a linear change in resonant frequency as the length of thetube changes. Note that this data was obtained without the perforated screen, thepresence of which will change the absolute values of the resonant frequencies. Anapproximate 14% change in the tube length resulted in a corresponding 14% change inthe second resonant frequency, which changed from approximately 1700 Hz to 1450 Hzover the full range of motion of 15 mm. From these characterization results, it isexpected that the adaptive length tube will work better than the adaptive-flap techniqueover a wide range of frequencies.

23

250 500 750 iO00 1250 [ 500 1750 2000

Figure 3.13: FRF showing adaptivedength HQ tu_ :resonator frequency change withlengnt_ parameter D.

Twenty tu_s were coastr'acted with the adaptive"length mechanism., i_stalled o_ theJTISDo Each Vabe was adjusted by a separate s_epper-motor dr/yen with the. same driverto obtain the _ame adj_tment in all _e tube:_. Figure 3, t4 eon_:ai_s a picture showir_g theadaptive array ialet moderated, on the JTISD _d a close-up of the adaptatk_r_ mec:hani.smwith the stepper motors,

Figure 3A4:, Pictures showirtg expe_mental contig,aration .fi_radapt:ive_Ie_gth HQ robeson the JTI 5D,

24

3.4

The adap_veqength HQ_tube system shown in figttre 3.14 was tested on file _I5Dengine. As an illustrative example of the performance of the system, the sound power

: 0 to 90 sector for three different rabe lengt_ at an engine speed$pectrttm over tile _ o ......yielding a BPF tone ti'equeney of 2250 Hz was obta_'ned. Figure 3.t5a-c show the soundpower sp_tntm tbr the bard wall condition arid for the adaptive at tube lengthsof t.1.5, 12o7, and 13.3 cm, respectively; It is dear that for the shortest tube length theopium broadband attenuation at the first and second tube resonance occurs near 1200and 2500 Hz, respectively. These frequencie_ of attenuation have s_Red towards lowerfrequencies a_ the tube length is increased from 11.5 to 133 cm. _ .fig_e 3A5c, thefrequencies of attenuation are now at approximately 10:50 and 2000 _, respectively.

The BPF tone _und power reductions for the three cases sho_,_ in the figures 3.lSa-eare 2,5, 5, I, and 5.2 dB, :respecfivdy. This result shows that the optim_ra length at theBPF tone of 2250 Hz is 13.3 era. It is thus e_pected that the optimum length tbr higherBPF tone frequencies would be shorter_ The effect of the tu_ length on the BPF tortepower reduction _II be demonstrated in the next section.

1t0

m

iO0

gsa. 90

O11

.......................................................................................Hard WailF_i ...................................L= tt,5cm

........................................L = 13,0 cm I

85 :..,

80 _i "

500 I000 i500 2000 2500 3000 ,3_0

....... HQ-tube for tube lengths ofFigure 3,15: Sotmd power sp.ec_ for adapuve-len_....(a)I 1.5, (_) 12_7, and (e) t3.3 era.

The performance and effect on the frequencies of attermation of the adaptive-len_system is probably easiest to _.ssess by plotting the broadband so_rM power reduction as afunction of the tube length. _s result will also be best to compare to the _tive-fl_mechanism. The sotmd power reduction as a fia._c.tioa of tube length is plotted in figure3.16a. It t_ er/clear that by changtag the length of the tubes, the frequency of power

25

attenuation around bo_h the first and seeped tube resonance .fi'e_ueneies change_ Imear|y.The dashed lines in the fig_e indicate the center _quenci.es of attenuation as a fim_tionof the tube l_ngth. The frequency changes for _e first and second tube, resonance areapproximately 300 and 600 Hz, -_spectively_ R _s into,resting to compare theof the two adaptation mechanisms. To _s end, the broadband sotmd power t_r the "ilap _"meehardsm i.s also plotted ir_. :figure 3.!6 (same as figure 310a)_ It is evident that the"flap" mechanism changes only the second tube resonance .frequency (ov_ the: anglesa.=25" m_d a_-55 _) as compared m "!ength" m_hani_m that affects both the tuberesonance .frequencies.

(b)Freq,(H_) dB Freq,{Hz)

t3,3 :3,(I :2,7 :2,4 _2,: : L8 I [,5

T_be Length (era)

Figure 3,16: Broa_thand setmd pressure level red:uctiort using adatstive (a) length and (b)flap HQ-tube _plement_tien.

3.5 ADAPTIVE CONTROL SYSTEM

Adaptive control can be used to adjus_ fl_e tube parameter to optimize theattenuation of the BPF tone over a rsa:ge of engin.e speeds. The te_ adap_ve control isused here imp!_in_ an ":actuator" is implemented m adjust the HQ_system based enerror reformat:on obtained from the sound rad_atmn from _he rater. Though this is in.

fact a_ active s?_atem, it Js not referred to as such m differemiate from previous activenoise control _proaehes where a secoltdary soulr_d field is introduced to destructivelyinterfere with the. fsx_. aoi_o In other engineering fields_ the adaptive system propose.here is eon_aonly referred to a_ _embaefive co_txol,

The adaptation of the tuBe'_ dynamics can be implemented with a control scheme asdepiet_ in the of figure 3.17. The error signal used by the control.let isob .rained from a_: _r_:ty of microphones that cart be mob:ted ie the f_x-field (su:h as onthe fitselage) or m the engine irdeL The signals t?rom ffvese micropt_mes are i_pu: imphigh_Q ba_@ass filter (BPF extractor)to extract _he signa_ _ the BPF fi_om _ total.

26

signal.SincetheBPFchangeswith time,aproximityprobefixedontheinlet nearthefanis usedto track the BPF frequency.TheBPFprobesignalis usedto adaptthecenterfrequencyof theband-passfilter. Thefilterederror signalsarethena purelysinusoidalsignal at the BPF. The RMS squaredvalue of the BPF signal is thenprocessedtogeneratea signalproportionalto thetotal acousticpowerradiatedby thefanat theBPFovera sectoron thefar-field,i.e. sidelines.Thesectorwherereductionis soughtshouldbeselectedto havethemaximumimpacton theEPNIAB metric.An actuatorwill thenadjustthetubepropertiesto minimizetheobservedradiatedpoweroverthesector.

The adaptive-lengthmechanismwasselectedfor the demonstrationof the adaptivecontrolsystemdescribedabove.Theerrormicrophoneswerepositionedon thefar-fieldat 18, 30 , and 42 to minimize the BPF tone radiated over this sector. The microphonelocations were selected because the inlet sound power is dominated by the flow ofacoustic energy over the 15 to 50 sector.

Proximity Adaptive

Inlet

AdapdveBPF

\\-3 Microphones

at 18c, 30 and42

!

II

I

Ir r_

AcousticPower

Estimator

A_uator toadapttubes

Figure 3.17: Schematic of adaptive control system.

The BPF extractor in figure 3.17 is needed to separate the BPF tone from thebroadband component before the power of the BPF tone is estimated. The approach usedhere to separate the BPF tone is based on using switched-capacitor filter and phase-lock-loop circuits as shown in figure 3.18. The BPF tone is obtained from the total signal byusing an analog, high-Q bandpass switched-capacitor filter whose center frequency isadjusted using an external clock signal. By changing the frequency of this clock signal,

27

thecenterfrequencyof thebandpassfilter canbechanged.To thisend,areferencesignalrelatedto the BPF toneis usedasan input into a Phase-Locked-Loopcircuit which isdesignedto generatetheclocksignal.Takingthereferencesignalfromthe fanproximityprobetransducer,thereferencesignalchangestheclocksignalandthusthe filter centerfrequencyastheenginespeedvaries.Therefore,thebandpassfilter is alwayscenteredattheBPFtone.TheBPFtoneis thenprocessedto obtainits power,i.e.mean-square-value.Thoughextractingthe BPF tone signal canalso be achievedusing adaptivedigitaltechniques[7], the main advantageis the very low costof the analogcomponentsascomparedto digital signalboards.This approachhasbeensuccessfullyimplementedonthecontrolof high-cycle-fatigueof engineblades[8].

_1 Switched _ Ir I Capacitor Filters I r I RMS Circuits yMicrophone I A I BPF BPF

Output IF Tone PowerCIock Signal

I Phase'Lcked II Loops I

BPF Signal t

Figure 3.18: BPF tone extractor system.

The performance of the BPF tone extractor is demonstrated in figures 3.19 and 3.20.The time histories of one of the original microphone signals and the BPF tone component(after the BPF extractor) are shown in figure 3.19. The corresponding frequency spectraare shown in Figure 3.20. This figure clearly indicates that only the BPF tone signal isfiltered out while the broadband component is almost completely rejected. These resultsclearly demonstrate the very good performance of the analog BPF extractor circuit.

28

(a)

"8

-_-'0

o4

_80 025 0.5 0.75 : 1.25 1,5

The BPF tone components from the microphones were then used to estimate theacoustic power radiated over the sector in the far-field where the error microphones werelocated. The microphone BPF tone signals were digitized and the mean-square-valuesestimated using a moving average approach. The mean-square-value signals were thenadded to give an estimate of the sound power, i.e. error signal to be minimized.

To investigate the effect of the tube length over the BPF tone power estimate, theerror signal was plotted as a function of the tube length at three engine speeds resulting inBPF tone frequencies of 2225, 2350, and 2400 Hz, respectively. The sound powerradiated over the selected sector in the far-field (error signal) as a continuous function ofthe tube length is plotted in figures 3.21a-c for the three BPF tone frequencies. Todetermine the attenuation as a function of the tube length, the error signal was alsorecorded for the hard wall case for the same three BPF tone frequencies. The hard wallerror signal is indicated by the red line in figures 3.21a-c and it is obviously not afunction of tube length. It is important to remark that the error signal was not calibratedand thus the levels are not indicative of actual sound power levels. However, thereduction of the error signal is the reduction in sound power radiated over the monitoredsector.

It is very clear in these figures that there exits an optimum tube length for the systemat each BPF tone. It is interesting to observe that for the BPF tone frequency of 2350 Hzthere is a relatively narrow range of tube lengths that leads to the best attenuation. Incontrast the other two frequencies show that the range of tube length that results in goodBPF tone reduction is not as sensitive as before. It is possible that the behavior at theextreme frequencies shown is because the true optimum is outside the range of tubelength implemented here.

The next experiments consisted of demonstrating the performance of the adaptivecontrol concept. To this end, a simple control algorithm was developed that changed thetube length to minimize the error signal, e.g. BPF tone power estimate. The controlalgorithm simply searched the minimum of the error signal (cost function) by evaluatingits slope (gradient search technique).

The experimental results from the adaptive control test are presented in figures 3.22a-c. The JT15D engine was started and set to operate such that the engine speed was around2400 Hz. Then the engine speed was reduced until the BPF tone frequency was down toabout 2250 Hz over several minutes. During this time, the error signal, tube length, andBPF tone frequency time histories were recorded and are shown in figures 3.22a through3.22c. Figure 3.22c shows that over the first 60 seconds, the engine speed was keptapproximately constant to yield a BPF tone of 2400 Hz. The initial tube length was set at12.6 cm which is not close to the optimum length for this frequency (see figure 3.21). It isclear from figure 3.22b that the controller adjusted the tube length to the optimum valueof 11.1 cm in about 40 seconds. Soon after the controller adjusted the tube length, theengine speed was slowly decreased as shown in figure 3.22c. The controller kept

adjusting the HQ-system by increasing the tube length. The symbol (o) in figure 3.22indicates an estimate of the optimum tube length at the corresponding BPF tone

3O

frequencyobtainedfromfigures3.2la-c.

Finally the performanceof the adaptive HQ-system is validated from the results infigure 3.22a. In this figure, the time history of the error signal is shown for the adaptiveHQ-system. To determine the attenuation of the error signal, the error signal for the hardwall case as a function of the BPF tone frequency is plotted as a continuous red line. It isclear that the adaptive system was capable of maintaining a reduction of more than 6 dBon the error signal, and thus of the radiated power over the sector where the errormicrophones are located, over most of the 6 minutes of testing time.

The results from this adaptive HQ-system experiment clearly demonstrate thepotential benefits of adjusting the HQ-properties to optimize its performance over a rangeof frequencies. The control system it has a very slow response time. However, the controlalgorithm selected for this first set of experiments was very simple and can be improved.In addition, a real implementation will probably not require a very fast time response forthe system.

31

(a)75

,70

65

6O

55 -

5O11

Hard Wall lnlet Level ~ 69" 7 dB t

BPF= 2225 Hz

11,5 12Tube Length (em)

12.5 13

(b)

(c)

75

7O

-_ 65

60

55

50 ' ' '11

Hard Walt Inlet Level _ 68.5 dB

,t |tB

i " _'_

; = _., "" " BPF = 2350 I_

..... o,._lt _ _ _ _ [ , , _, I11.5 12 12.5 13

Tube Length (cm)

75

70

65

60

55

5011

L t It_llll

Ha_ Wall _ L_I ~ _ dB

11.5 12 12.5 13Tube Length (cm)

Figure 3.21: Estimated power by error-microphones as a function of tube length at(a) 2225, (b) 2350, and (c) 2400 Hz.

32

7I55 Zar__"v

5o55

_ 5o Adaptive HQ-Tube lnlet45

0 60 120 180 240 300 360

13

11.5

11 '

0 60 120 180 240 300 360

2450

2400

"_' 2350

=_23002250

\

(c)

\

22O0 _ i i _ _ I i i t I t I i _ r i i t i i _ i i I i i i i i I i i i i i0 60 120 180 240 300 360

33

4. CONCLUSIONS

In general, the experiments performed on the JT15D demonstrate that the applicationof HQ tubes to the problem of turbofan jet engine noise is a very effective and viablestrategy. In this section, the main conclusions obtained from the research performed inthis experimental part of the research endeavor are discussed.

The first part of this report showed the experimental results obtained with one andtwo arrays of Herschel-Quincke (HQ) tubes installed on the JT15D engine inlet, with theengine configured without excitation rods. All previous tests with the HQ tubes wereperformed with 27 rods located upstream of the fan [1] that resulted in the m=lcircumferential order mode to dominate the inlet sound field. These tests performedwithout the rods resulted in the m=5 circumferential order mode to dominate the inletsound field. Both broadband and tonal data were obtained for a hard-walled inlet and for

the inlet configured with one and two arrays of HQ tubes. The good attenuation of thebroadband and BPF tone components once again demonstrated the potential of the HQ-concept.

The second part of the report described the efforts to develop adaptive capabilitiesfor the HQ-system. A preliminary investigation into the potential of using adaptive HQtubes was presented. It is clear that although the tubes are very effective over a range ofBPFs, there is an optimal BPF at which the best attenuation occurs. This implies that inorder to optimize the attenuation effect of the tubes over a wide range of BPF tones, thetubes should be adapted to "track" the BPF as it changes. Two adaptation mechanismswere proposed and built. One approach consisted of using a "flap" placed at the center ofthe tube. It was shown that this flap changed the 2 nd resonance of the tube withoutaffecting the first one. The second approach consisted of changing the tube length usingexpandable elements. This technique demonstrated that adjusting the tube length shiftsboth the 1st and 2nd tube resonances. This last adaptation mechanism was alsoimplemented in conjunction with an adaptive control algorithm. Three microphones wereplaced in the far-field and used to estimate the power radiated over a sector. The controlalgorithm adjusted the tube length to minimize this power estimate as the engine speedwas changing. One of the main conclusions of the work on the adaptive-HQ concept isthat adaptation of the HQ-tubes to optimize performance is feasible. However, thefrequency range in which the adaptive system can operate is probably limited, e.g. 10%of the BPF tone. Thus, the potential application of an adaptive HQ-system might be inoptimizing the performance for a single engine power setting, i.e. approach, cutback, andso forth. It would be difficult to device an adaptive HQ-system that can be adjusted tooptimally work at more than one power setting.

5. RECOMMENDATIONS FOR FUTURE RESEARCH

The adaptive HQ-system is an approach that needs to be further investigated. One ofthe most important issues is to determine the practical conditions under which theadaptation concepts would be most beneficial. It is obvious that much work is needed in

34

practicaladaptationmechanisms,in particularfor mechanismsthatdonotrequiremovingparts.Someoptionsincludeaffectingthetubeacousticsthroughflow injection(similartobias-flowin liners)or temperaturegradients.

Anotherareaof researchfor the implementationof adaptiveHQ-systemsis in thedevelopmentof sensingconceptsandcontrol algorithms.The developmentof controlalgorithmsis probablynot critical sincethere arenumeroustechniquesthat canbeimplementedin this application.However,the sensingconceptrequiresfurtherresearchefforts.To adjustthe tubeproperties,it is importantto obtain"error" informationthatrelatesto the inlet radiatedacousticpower.Onefeasibleapproachis to placeaxialarraysof microphoneson the inlet wall to estimatetheradiatedpowerover far-field sectorsusingwavenumberconcepts[9,10].

ACKNOWLEDGEMENTS

This work was supported by the Aeroacoustics Branch of the NASA LangleyResearch Center which is gratefully acknowledged. The technical monitors for this workare Dr. Carl Gerhold and Dr. Joe Posey.

REFERENCES

[1] Smith, J.P. and Burdisso, R. A., "The Application of the Herschel-Quincke TubeConcept for the Reduction of Tonal and Broadband Noise From Turbofan Engines,"VPI report VPI-ENGR.98.167, prepared for NASA under grant # NAG-l-1980 andproposal # 98-0448-10, 1998.

[2] Preisser, J. S., Schoenster, J. A., Golub, R. A., and Home, C., "Unsteady Fan BladePressure and Acoustic Radiation From a JT15D-1 Turbofan Engine at SimulatedForward Speed," AIAA Paper 81-0096 presented at the 19th Aerospace SciencesMeeting, January 12-15, 1981.

[3] Burdisso, R. A., Thomas, R. H., Fuller, C. R., and O'Brien, W. F., "Active Controlof Spinning Modes from a Turbofan Engine," AIAA Journal, Vol. 32, No. 1, pp.23-30, 1994.

[4] Burdisso, R.A., Fuller, C.R., Smith, J.P., "Experiments on the Active Control of aTurbofan Inlet Noise using Compact, Lightweight Inlet Control and ErrorTransducers," CEAS/AIAA-95-028, 1995, pp. 177-185.

[5] Smith, J.P., Burdisso, R.A., and Fuller, C.R., Experiments on the Active Control ofInlet Noise From a Turbofan Jet Engine Using Multiple Circumferential ControlArrays, AIAA 96-1792, 1996.

35

[6]

[71

[8]

[9]

[10]

Homyak, L., McArdle, J. G., and Heidelberg, L. J., "A Compact Inflow ControlDevice for Simulating Flight Fan Noise," AIAA Paper no. 83-0680 presented at theAIAA 8tnAeroacoustic Conference, April 11-13, 1983.

Widrow B. ans Steams S. D., "Adaptive Signal Processing,", Prentice-Hall SignalProcessing Series, Englewood Cliffs, N.J., 1985.

J. Feng, J. Kozak, T. Bailie, R. A. Burdisso and W. N. Ng, "Active Flow ControlUsing Microphone Sensors to Reduce Fan Noise," AIAA-2000-1993, 6 thAIAA/CEAS Aeroacoustics Conference, Maul, Hawaii, USA, 12-14 June 2000.

Smith J. P. and Burdisso R. A., "Active Control of Inlet Noise from a TurbofanEngine Using Inlet Wavenumber Sensors," 5th AIAA/CEAS AeroacousticsConference AIAA 99-1808, Seattle, Washington, USA, 10-12 June 1999.

Smith J. P., Burdisso R. A., Sutliff D.L., and Heilderberg L. J. "Active Control on aLarge-scale Ducted Fan Inlet With Wavenumber Sensing," 139th Meeting of theAcoustical Society of America, Atlanta, Georgia, May 31 2000.

36

Form ApprovedREPORT DOCUMENTATION PAGE OMBNo.0704-0188

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources,gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of thiscollection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 JeffersonDavis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503.1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORTTYPE AND DATES COVERED

March 2002 Contractor Report4. TITLE AND SUBTITLE 5. FUNDING NUMBERS

Experiments With Fixed and Adaptive Herschel-Quincke Waveguides on thePratt and Whitney JT 15D Engine

6. AUTHOR(S)Jerome P. SmithRicardo A. Burdisso

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)Virginia Polytechnic Institute and State UniversityDepartment of Mechanical EngineeringBlacksburg, VA 24061

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)

National Aeronautics and Space AdministrationLangley Research CenterHampton, VA 23681-2199

G NAG1-2137

WU 706-81-12-01

8. PERFORMING ORGANIZATIONREPORT NUMBER

VPI-ENGR 4-26483

10. SPONSORING/MONITORINGAGENCY REPORT NUMBER

NASA/CR-2002-211430

11. SUPPLEMENTARY NOTES

NASA Langley Technical Monitor: Carl H. Gerhold

12a. DISTRIBUTION/AVAILABILITY STATEMENT

Unclassified-UnlimitedSubject Category 71 Distribution: NonstandardAvailability: NASA CASI (301) 621-0390

12b. DISTRIBUTION CODE

13. ABSTRACT (Maximum 200 words)This report summarizes the key results obtained by the Vibration and Acoustics Laboratories at Virginia Tech overthe period from January 1999 to December 2000 on the project "Investigation of an Adaptive Herschel-QuinckeTube Concept for the Reduction of Tonal and Broadband Noise from Turbofan Engines", funded by NASA LangleyResearch Center. The Herschel-Quincke (HQ) tube concept is a developing technique the consists of circumferen-tial arrays of tubes around the duct. A fixed array of tubes is installed on the inlet duct of the JT15D research engineat Virginia Tech. In addition, an adaptive configuration is evaluated in which the H-Q noise control system is opti-mized by changing the parameters of the tubes as the engine speed is varied.

14. SUBJECTTERMSduct propagation, experiment, turbofan, Herschel-Quincke tube

15. NUMBER OF PAGES4O

16. PRICE CODE

17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATIONOF REPORT OFTHIS PAGE OF ABSTRACT OF ABSTRACT

Unclassified Unclassified Unclassified UL

NSN 7540-01-280-5500 Standard Form 298 (Rev. 2-89)Prescribed by ANSI Std. Z39-18298-102